THE PERFORMANCE and ASSET ALLOCATION of German ROBO-ADVISORS

Abstract

After a short historical perspective on the emergence of robo-advisors and an overview of how they manage other people's money, we evaluate the performance of fi ve German robo-advisors in the period between May 2015 and December 2018. Performance tests are conducted using Sharpe's (1966) and Jensen's (1968) performance methodologies. We also employ the returns-based style analysis of Sharpe (1992) to determine the exposure of robo-advisors to different non-overlapping asset classes. We report the following fi ndings: First, no robo-advisor was able to beat the benchmark before or after considering fees. Second, robo-advisor performance varies greatly in the sample period even for portfolios that should appeal to clients with similar risk preferences. Third, these performance differences remain unexplained after accounting for the different asset allocations.